caffe 使用cudnn 加速报错

./include/caffe/util/cudnn.hpp: In function ‘void caffe::cudnn::createPoolingDesc(cudnnPoolingStruct**, caffe::PoolingParameter_PoolMethod, cudnnPoolingMode_t*, int, int, int, int, int, int)’:
./include/caffe/util/cudnn.hpp:145:3: error: ‘CUDNN_PROPAGATE_NAN’ was not declared in this scope
CUDNN_CHECK(cudnnSetPooling2dDescriptor_v4(*pool_desc, *mode,
^
./include/caffe/util/cudnn.hpp:145:3: error: there are no arguments to ‘cudnnSetPooling2dDescriptor_v4’ that depend on a template parameter, so a declaration of ‘cudnnSetPooling2dDescriptor_v4’ must be available [-fpermissive]
CUDNN_CHECK(cudnnSetPooling2dDescriptor_v4(*pool_desc, *mode,
^
./include/caffe/util/cudnn.hpp:145:3: note: (if you use ‘-fpermissive’, G++ will accept your code, but allowing the use of an undeclared name is deprecated)
./include/caffe/util/cudnn.hpp: At global scope:
./include/caffe/util/cudnn.hpp:151:40: error: variable or field ‘createActivationDescriptor’ declared void
inline void createActivationDescriptor(cudnnActivationDescriptor_t* activ_desc,
^
./include/caffe/util/cudnn.hpp:151:40: error: ‘cudnnActivationDescriptor_t’ was not declared in this scope
./include/caffe/util/cudnn.hpp:151:69: error: ‘activ_desc’ was not declared in this scope
inline void createActivationDescriptor(cudnnActivationDescriptor_t* activ_desc,
^
./include/caffe/util/cudnn.hpp:152:27: error: expected primary-expression before ‘mode’
cudnnActivationMode_t mode) {
^

make: *** [.build_release/src/caffe/util/db.o] Error 1

解决方法：

因为ssd 太老了，我的cudnn 是5.1 的版本

所以需要两者相匹配：

caffe 使用cudnn 加速报错相关推荐

关于PR硬件加速报错错误代码3解决办法
1.设备管理器禁用集显 2.PR 编辑-首选项-媒体硬件加速编码,取消INTER的✓.保存重启PR. 如果这个不行.用下面这个方法把PR设置成管理员运行.
Deep Learning for Computer Vision with Caffe and cuDNN
转载自:Deep Learning for Computer Vision with Caffe and cuDNN | Parallel Forall http://devblogs.nvidia. ...
./include/caffe/util/cudnn.hpp: error: too few arguments to function ‘SetPooling2dDescriptor
https://blog.csdn.net/tianrolin/article/details/71246472 这是由于所使用的cuDNN版本不一致的导致的,作者配置环境是cuDNN 4.0,但是5 ...
深度学习自学（十六）：caffe-sphereface识别代码编译问题－caffe sudo make test报错
一.caffe-sphereface编译错误 caffe sudo make test报错 sudo make runtest -j33 CXX src/caffe/test/test_benchma ...
Ubuntu14.04 + Matlab2014a + caffe + cuda + cudnn环境搭建
为什么80%的码农都做不了架构师?>>> 安装Caffe环境前,请查好各个软件版本的相互匹配资料,确定好所有软件版本后,再开始安装. 一.安装gcc4.6 Ubuntu14.0 ...
caffe 加入 cudnn编译
首先要使用cudnn v2, ALL PLATFORMS Extract the cuDNN archive to a directory of your choice, referred to be ...
Rsa2加密报错java.security.spec.InvalidKeySpecException的解决办法
最近在和支付宝支付做个对接,Java项目中用到了RSA2进行加解密,在加密过程中遇到了错误: java.security.spec.InvalidKeySpecException: java.secu ...
caffe运行训练脚本时报错：Unknown bottom blob 'data' (layer 'conv1',bottom index 0)
报错的两种报错原因: 1.输入数的路径错误,需要将路径进行修改排查目录是否出错 2.训练原数据格式不对 3.train.prototxt文件中并未设置test层,而在solver层则设置了test的迭 ...
arduino编程时加{}报错_使用Arduino开发板时最常见的10个错误
作为Arduino开发板的初学者,对于没有电子背景的人来说非常具有挑战性,你会遇到很多错误,其中一些可能有简单的解决方案,但可能需要几天的时间来解决.因此,为了使事情变得更容易,我制作了一个包含10个 ...

caffe 使用cudnn 加速报错

caffe 使用cudnn 加速报错相关推荐

最新文章

热门文章